Cellular immunities, particularly cytotoxic T cells (referred to as CTLs hereinafter), play an important role in the elimination of cancer cells or virus-infected cells from a living body. CTLs recognize a complex formed between an antigen peptide derived from a cancer antigen protein on a cancer cell (cancer antigen peptide) and an MHC (Major Histocompatibility Complex) class I antigen (referred to as an HLA antigen in the case of human), and thereby attack and injure cancer cells.
Representative examples of cancer antigen proteins are listed in Table 1 described in Immunity, vol.10: 281, 1999. Specific examples include melanosomal antigens such as a melanocytic tissue-specific protein, gp100 (J. Exp. Med., 179:1005, 1994), MART-1 (Proc. Natl. Acad. Sci. USA, 91:3515, 1994), and tyrosinase (J. Exp. Med., 178:489, 1993); as well as HER2-neu (J. Exp. Med., 181:2109, 1995) and cancer markers such as CEA (J. Natl. Cancer Inst., 87:982, 1995) and PSA (J. Natl. Cancer Inst., 89:293, 1997) as cancer antigen proteins other than those from melanomas. Cancer antigen peptides are peptides consisting of about 8 to 11 amino acid residues, generated through the processing of cancer antigen proteins with intracellular proteases (Cur. Opin, Immunol., 5:709, 1993; Cur. Opin, Immunol., 5: 719, 1993; Cell, 82: 13, 1995; Immunol. Rev., 146: 167, 1995). The cancer antigen peptides thus generated bind to MHC class I antigens (HLA antigens) to form complexes, and then the complexes are presented on cellular surfaces, and recognized by CTLs as described above. In development of medicaments for cancer immunotherapy (cancer vaccines) based on cancer cells disruption by CTLs, it therefore is very important to identify a cancer antigen peptide from the cancer antigen protein, which can effectively induce CTLs.
Lots of subtypes exist in MHC class I molecules, and the amino acid sequence of an antigen peptide that binds to the respective subtype obeys a certain rule (binding motif). Regarding the binding motif for HLA-A2, for example, the amino acid at position 2 is leucine, methionine, or isoleucine, and the amino acid at position 9 is valine, leucine, or isoleucine. Regarding the binding motif for HLA-A24, the amino acid at position 2 is tyrosine, phenylalanine, methionine, or tryptophan, and the amino acid at position 9 is phenylalanine, leucine, isoleucine, tryptophan, or methionine. Recently, any peptide sequence expected to be capable of binding to HLA antigens including the motifs as shown above may be searched on databases (for example, BIMAS software; http://bimas.dcrt.nih.gov/molbio/hla_bind/). Accordingly, in order to identify a cancer antigen peptide that can induce CTLs from the cancer antigen protein, peptide regions consisting of about 8 to 11 amino acid in length that match the binding motif or the peptide sequence expected for an intended HLA type are first identified from the amino acid sequence of the cancer antigen protein.
However, peptides that have been identified based on the binding motif or the expected peptide sequence are not necessarily immunogenic. Since an antigen peptide is generated through the intracellular processing of a cancer antigen protein, a peptide not having been generated through the processing cannot be an antigen peptide. Furthermore, since many cancer antigen proteins exist originally in a living body, CTLs may be tolerant to such cancer antigens even if a peptide having the binding motif or the expected peptide sequence is intracellularly generated as a cancer antigen peptide. Those show that, in order to identify a cancer antigen peptide having an activity to induce CTLs, a prediction merely based on the binding motif or the peptide sequence expected for an intended HLA type is insufficient, and an in vivo evaluation for immunogenicity (an activity to induce CTLs) should be important.
A Wilms cancer suppressor gene WT1 (WT1 gene) was isolated from chromosome 11p13 as one of the causative genes of Wilms cancers based on the analysis of the WAGR syndrome that was complicated by Wilms cancers, aniridia, urogenital anomaly, mental retardation, etc. (Nature, 343: 774, 1990). The genomic DNA of WT1 is about 50 Kb, and is composed of ten exons, of which cDNA is about 3 kb. The amino acid sequence deduced from the cDNA is as shown in SEQ ID NO: 1 (Cell., 60:509, 1990). The WT1 gene has been suggested to promote the growth of leukemia cells from the facts that the WT1 gene is highly expressed in human leukemia, and that the leukemia cells are suppressed in their cellular growth by the treatment with WT1 antisense oligomers (Japanese Patent Publication (Kokai) No. 104627/1997). Then, the WT1 gene has been demonstrated to be a new cancer antigen protein of leukemia and solid cancers (J. Immunol., 164: 1873-80, 2000, J. Clin. Immunol., 20, 195-202, 2000) from the fact that the WT1 gene is also highly expressed in solid cancers such as gastric cancer, colon cancer, lung cancer, breast cancer, embryonal cancer, skin cancer, bladder cancer, prostate cancer, uterine cancer, cervical cancer, and ovarian cancer (Japanese Patent Publication (Kokai) No. 104627/1997, Japanese Patent Publication (Kokai) No. 35484/1999). Medicaments for cancer immunotherapy (cancer vaccines) can be preferably applied to as many as possible of cancer patients, and therefore it is important to identify cancer antigen peptides from WT1, which is highly expressed in many kinds of cancer, and to develop cancer vaccines based on those cancer antigen peptides. In this context, WO00/06602 and WO00/18795 describe naturally-occurring cancer antigen peptides composed of a portion of the WT1 protein.
In the course of development of cancer vaccines, evaluation of a vaccine for its in vivo efficacy cannot be conducted using pure-line mice commonly used as experimental animals, and requires an animal model for human expressing an HLA. Specifically, human antigen peptides usable as cancer vaccine induce specific immune responses when presented to an HLA, which is an MHC class I molecule specific for human. Non-human experimental animals lack such an HLA, and therefore are unavailable for in vivo evaluation of cancer vaccines directed to treatment of human. Accordingly, animal models for human expressing an HLA are essential in the evaluation of cancer vaccines for their efficacy as described above.